"Cognitive science" is not the name of any well-defined research project but rather a whole family of overlapping research areas, including work in what would traditionally be thought of as psychology, artificial intelligence (AI), linguistics, anthropology, philosophy and even neurophysiology. Nonetheless, within this vast stew of different research aims and methods, there was until fairly recently a prevailing core idea of cognitive science. To put it very crudely, the idea was that the mind functions like a digital computer. Mental states are computational states and mental processes are computational processes. Many hardcore cognitive scientists still insist that the "computer metaphor" is not a metaphor at all, that the brain is literally a digital computer and the mind is simply a computer program (see, for example, Pylyshyn 1984). For this reason, many researchers see work in AI as absolutely essential to the cognitive science enterprise. So in this chapter, when I talk about the prospects for cognitive science, I am going to be talking about the prospects for a research programme based on the artificial intelligence, computational model of the mind.²